The odderon intercept is calculated directly using the wave function recently constructed by R.A.Janik and J.Wosiek. The results confirm their reported value. It is also found that their solution for $q_3=0$ does not satisfy the Bose-symmetry requirements. Introduction of terms into the trial wave function with an asymptotical behavior similar to the Janik-Wosiek wave functions does not seem to improve variational estimates significantly. The diffusion parameter is found to be of the order 0.6.